Ableton Forum

The meters seem backward on glue compressor.

I compress a signal on a 2:1 ratio, I get 5dB gain reduction
I move the ratio to 10:1 I get 1dB of gain reduction

I don't get it. Does yours do the same?

I don't get the same result using SSL's plugin. Moving to 10:1 yields more gain reduction.

whats happening?

I hadn't thought about this before, but I'll venture a guess. The documentation says the (non-user-adjustable) knee becomes sharper as the ratio increases. Unless I'm making a mental error here, it seems if the input signal is hitting below the threshold, this would actually cause the amount of compression and therefore gain reduction to decrease as the ratio increases, contrary to expectation.

Will be interested to see if anyone can confirm or disconfirm this.

Right about to do dinner, but I had a thought. Is the "glue" a different compressor? I'll futz wit this after BLT's.........

Andy, the developer of The Glue compressor, answers your question here, where he says that each ratio has a different threshold offset to make comparing the effect of different settings easier (apparently modeled from the original circuit):

https://www.kvraudio.com/forum/viewtopi ... start=2205

fishmonkey wrote: ↑

Sun May 14, 2023 12:17 am

Andy, the developer of The Glue compressor, answers your question here, where he says that each ratio has a different threshold offset to make comparing the effect of different settings easier (apparently modeled from the original circuit):

https://www.kvraudio.com/forum/viewtopi ... start=2205

So here's what I'm wondering now. Someone in that thread made the same guess I did (about the knee), and someone does here as well, and this thread includes both answers (knee and threshold offset), but the developer says threshold offset. So is it actually true that the knee varies based on ratio?

yes, the knee is softer at lower ratios, and hardens as you increase the ratio.

it is possible to set The Glue up such that it is doing more gain reduction at higher ratios, but it is very program and settings dependent. and, the combination of the softer knee plus threshold offset seems to affect the 2:1 ratio more than the 5:1 ratio when comparing to the 10:1 ratio.

in the end it's these design choices that give different compressors their own flavours...

fishmonkey wrote: ↑

Sun May 14, 2023 12:48 am

yes, the knee is softer at lower ratios, and hardens as you increase the ratio.

I'm sure you're right, but I'm just curious, is there a source for this other than the manual? (Maybe in that big long thread somewhere?) The manual does have some errors here and there, so I'm just curious as to whether there's another source. (No stake one way or the other - just wondering.)

[Edit: Looks like you expanded your post after I posted this, and it looks like maybe you have inside knowledge on the matter, so I'll withdraw my question ]

fishmonkey wrote: ↑

Sun May 14, 2023 12:17 am

Andy, the developer of The Glue compressor, answers your question here, where he says that each ratio has a different threshold offset to make comparing the effect of different settings easier (apparently modeled from the original circuit):

https://www.kvraudio.com/forum/viewtopi ... start=2205

Thanks for searching for that thread. It explains why I get the differing results when moving the ratio knob. But Andy does say its modelled on the behaviour of the actual unit, which is weird because I also have SSL's bus compressor, and it behaves differently, in a more expected manner. I'm assuming we are talking about the same bus compressor found on the SSL console?

But if Andy says its supposed to behave like that I wont worry any more. Thanks heaps for clearing that issue up for me.

the other version of the SSL buss compressor i have is the PA one, bx_townhouse. it behaves similarly to The Glue in the 5:1 and 10:1 settings, but seems to have a smaller offset in the 2:1 setting, as the gain reduction tends to be a little lower at 2:1.

i agree that it is a little strange that the 2:1 setting in The Glue tends to hit significantly harder than the 5:1 or 10:1 settings. i'm guessing that this was a design choice based upon level-matching with various kinds of source material, however i can't be bothered to test that theory!

fishmonkey wrote: ↑

Sun May 14, 2023 12:17 am

Andy, the developer of The Glue compressor, answers your question here, where he says that each ratio has a different threshold offset to make comparing the effect of different settings easier (apparently modeled from the original circuit):

https://www.kvraudio.com/forum/viewtopi ... start=2205

This!

So futzing isn't going to yield much in the way of results..........

Seems to me, that ALL compressors, and maybe bus compressors in particular have unique variables that make them......unique?

bishbo2000 wrote: ↑

Sat May 13, 2023 9:19 pm

The meters seem backward on glue compressor.

I compress a signal on a 2:1 ratio, I get 5dB gain reduction
I move the ratio to 10:1 I get 1dB of gain reduction

I don't get it. Does yours do the same?

I don't get the same result using SSL's plugin. Moving to 10:1 yields more gain reduction.

whats happening?

The Bus and channel compressor seem very different from each other as one is feed forward and other is feedback.

This response seems the same in Reason, Ik, UAD, NI, or about every SSL type compressor I own. They just work differently and have different purposes.

jlgrimes wrote: ↑

Sun May 14, 2023 11:32 pm

The Bus and channel compressor seem very different from each other as one is feed forward and other is feedback.

This response seems the same in Reason, Ik, UAD, NI, or about every SSL type compressor I own. They just work differently and have different purposes.

interesting. i only have Plugin Alliance SSL stuff, and none of the ones i have show the behaviour that The Glue does, including SSL 4000 E/G, SSL 9000 J or the bx_townhouse...

fishmonkey wrote: ↑

Mon May 15, 2023 12:29 am

jlgrimes wrote: ↑

Sun May 14, 2023 11:32 pm

The Bus and channel compressor seem very different from each other as one is feed forward and other is feedback.

This response seems the same in Reason, Ik, UAD, NI, or about every SSL type compressor I own. They just work differently and have different purposes.

interesting. i only have Plugin Alliance SSL stuff, and none of the ones i have show the behaviour that The Glue does, including SSL 4000 E/G, SSL 9000 J or the bx_townhouse...

I probably just might not be paying enough attention but I have never had any issues with the Gain reduction indicators on the Glue or any SSL style bus compressor I've used.

I always have a problem more with the SSL style channel compressors (which Ableton doesn't have), where I have to lower the threshold considerably to get any "meaningful" gain reduction and this is common on about any SSL style channel compressor I've used. I was told that this is common on the hardware channel strips and that the gain reduction on those compressors are crude and don't show any gain reduction between 0 and 3db and I either have to drive the Input gain or use a very high ratio. The Bus Compressors (or any I've worked with) don't work that way at all (which I like)

coincidentally Dan Worrall analyses and discusses The Glue's ratio/threshold interaction in his most recent video:

https://youtu.be/1IH6N63AXu4?t=3268

in the Plugin Doctor analysis you can see that at the 4:1 setting the threshold knob reading matches the internal threshold of the compressor, and that there is a large threshold reduction in the 2:1 setting...